Usage question for interpolate_na

[alessioarena]

Hi All!

Just wondering what is the difference between the argument max_gap and limit in interpolate_na?

From the documentation:
max_gap - Maximum size of gap, a continuous sequence of NaNs, that will be filled
limit - Maximum number of consecutive NaNs to fill

http://xarray.pydata.org/en/stable/generated/xarray.DataArray.interpolate_na.html

[keewis]

consider this:

In [2]: da = xr.DataArray(
   ...:     [0, np.nan, 1, np.nan, np.nan, 2, np.nan, np.nan, np.nan, 3],
   ...:     dims="x",
   ...:     coords={"x": np.arange(-10, 10, 2)},
   ...: )
   ...: da
Out[2]: 
<xarray.DataArray (x: 10)>
array([ 0., nan,  1., nan, nan,  2., nan, nan, nan,  3.])
Coordinates:
  * x        (x) int64 -10 -8 -6 -4 -2 0 2 4 6 8

In [3]: da.interpolate_na(dim="x", method="nearest", limit=2)
Out[3]: 
<xarray.DataArray (x: 10)>
array([ 0.,  0.,  1.,  1.,  2.,  2.,  2.,  2., nan,  3.])
Coordinates:
  * x        (x) int64 -10 -8 -6 -4 -2 0 2 4 6 8

In [4]: da.interpolate_na(dim="x", method="nearest", max_gap=7)
Out[4]: 
<xarray.DataArray (x: 10)>
array([ 0.,  0.,  1.,  1.,  2.,  2., nan, nan, nan,  3.])
Coordinates:
  * x        (x) int64 -10 -8 -6 -4 -2 0 2 4 6 8

which means that:

• limit constrains the number of consecutive missing values filled per gap (after which it stops)
• max_gap is the maximum size of the gap (in coordinate space) that should be filled. Any gap bigger than that will not be filled at all.

I agree that it's pretty difficult to figure that out reading just the parameter description. Maybe examples demonstrating that would make it easier to understand? If you're up for it, we would definitely welcome a PR adding those.

[dcherian]

👍 for adding examples.

limit is copied from pandas. I added max_gap because I never found limit's behaviour to be useful (pandas still has outstanding issues asking for max_gap).